from diffusers import StableDiffusionPipeline from diffusers import StableDiffusionImg2ImgPipeline import gradio as gr import torch from PIL import Image import utils is_colab = utils.is_google_colab() max_width = 832 max_height = 832 class Model: def __init__(self, name, path, prefix): self.name = name self.path = path self.prefix = prefix models = [ Model("Custom model", "", ""), Model("Arcane", "nitrosocke/Arcane-Diffusion", "arcane style "), Model("Archer", "nitrosocke/archer-diffusion", "archer style "), Model("Elden Ring", "nitrosocke/elden-ring-diffusion", "elden ring style "), Model("Spider-Verse", "nitrosocke/spider-verse-diffusion", "spiderverse style "), Model("Modern Disney", "nitrosocke/modern-disney-diffusion", "modern disney style "), Model("Classic Disney", "nitrosocke/classic-anim-diffusion", ""), Model("Waifu", "hakurei/waifu-diffusion", ""), Model("Pokémon", "lambdalabs/sd-pokemon-diffusers", ""), Model("Fuyuko Waifu", "yuk/fuyuko-waifu-diffusion", ""), Model("Pony Diffusion", "AstraliteHeart/pony-diffusion", ""), Model("Robo Diffusion", "nousr/robo-diffusion", ""), Model("Cyberpunk Anime", "DGSpitzer/Cyberpunk-Anime-Diffusion", "dgs illustration style "), Model("Hergé Style", "sd-dreambooth-library/herge-style", "herge_style "), ] current_model = models[1] current_model_path = current_model.path pipe = StableDiffusionPipeline.from_pretrained(current_model.path, torch_dtype=torch.float16) if torch.cuda.is_available(): pipe = pipe.to("cuda") device = "GPU 🔥" if torch.cuda.is_available() else "CPU 🥶" def custom_model_changed(path): models[0].path = path current_model = models[0] return models[0].path def inference(model_name, prompt, guidance, steps, width=512, height=512, seed=0, img=None, strength=0.5, neg_prompt=""): global current_model for model in models: if model.name == model_name: current_model = model model_path = current_model.path generator = torch.Generator('cuda').manual_seed(seed) if seed != 0 else None if img is not None: return img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, width, height, generator) else: return txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, generator) def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, generator=None): global pipe global current_model_path if model_path != current_model_path: current_model_path = model_path pipe = StableDiffusionPipeline.from_pretrained(current_model_path, torch_dtype=torch.float16) if torch.cuda.is_available(): pipe = pipe.to("cuda") prompt = current_model.prefix + prompt results = pipe( prompt, negative_prompt=neg_prompt, num_inference_steps=int(steps), guidance_scale=guidance, width=width, height=height, generator=generator) image = results.images[0] if not results.nsfw_content_detected[0] else Image.open("nsfw.png") return image def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, width, height, generator): global pipe global current_model_path if model_path != current_model_path: current_model_path = model_path pipe = StableDiffusionImg2ImgPipeline.from_pretrained(current_model_path, torch_dtype=torch.float16) if torch.cuda.is_available(): pipe = pipe.to("cuda") prompt = current_model.prefix + prompt ratio = min(max_height / img.height, max_width / img.width) img = img.resize((int(img.width * ratio), int(img.height * ratio))) results = pipe( prompt, negative_prompt=neg_prompt, init_image=img, num_inference_steps=int(steps), strength=strength, guidance_scale=guidance, width=width, height=height, generator=generator) image = results.images[0] if not results.nsfw_content_detected[0] else Image.open("nsfw.png") return image css = """ """ with gr.Blocks(css=css) as demo: gr.HTML( f"""
Demo for multiple fine-tuned Stable Diffusion models, trained on different styles:
Arcane, Archer, Elden Ring, Spiderverse, Modern Disney, Waifu, Pokemon, Fuyuko Waifu, Pony, Hergé (Tintin), Robo, Cyberpunk Anime
Running on {device}